Mailing-List: contact user-help@hive.apache.org; run by ezmlm
Precedence: bulk
Reply-To: user@hive.apache.org
Received-SPF: pass (athena.apache.org: domain of java8964@hotmail.com
 designates 65.55.116.91 as permitted sender)
Message-ID: <BLU162-W55258E8D496684D356F7BBD0480@phx.gbl>
Content-Type: multipart/alternative;
	boundary="_c0158b0f-e0e0-4d70-9531-283d71becc77_"
From: java8964 java8964 <java8964@hotmail.com>
To: <user@hive.apache.org>
Subject: Array index support non-constant expresssion
Date: Tue, 11 Dec 2012 17:24:52 -0500
Importance: Normal
In-Reply-To: 
 <CAKmbbzWBjN1JaRqButJNAkkN6SZe9ZRc7jY-EAEt-3Om-2Y94A@mail.gmail.com>
References: 
 <CAKmbbzWBjN1JaRqButJNAkkN6SZe9ZRc7jY-EAEt-3Om-2Y94A@mail.gmail.com>
MIME-Version: 1.0

--_c0158b0f-e0e0-4d70-9531-283d71becc77_
Content-Type: text/plain; charset="iso-8859-1"
Content-Transfer-Encoding: quoted-printable


Hi=2C In our project to use the HIVE on CDH3U4 release (Hive 0.7.1)=2C I ha=
ve a hive table like the following:
Table foo (   search_results array<struct<id:bigint=2C..............>>   se=
arch_clicks array<struct<index_loc:int=2C ..............>>)
As you can see=2C the 2nd column=2C which represents the list of search res=
ults clicked=2C contains the index location of which results was clicked (s=
tarting from 0). Now I need to find out the click count of some IDs=2C but =
this IDs only exists in the search_result struct.
When I tried to write a query for this=2C I will try to do following first:

select search_results[c.index_loc] from foo lateral view explode(search_cli=
cks) clickTable as c=3B
But it won't work in hive=2C as the following error message returned:AILED:=
 Error in semantic analysis: Line 1:7 Non-constant expressions for array in=
dexes not supported rank.
It looks like that currently Hive (at least 0.7) does NOT support non-const=
ant expressions as the array index.
I searched on google=2C and found out the following HIVE jira ticket:
https://issues.apache.org/jira/browse/HIVE-1955
It looks like someone has the same request to support it=2C but not yet.
But there is a comment in the above ticket that it can be done in an UDF. M=
y question is=2C can anyone share some ideas about how to archive this in a=
n UDF=2C as it maybe the only option for me right now?
Thanks 		 	   		  =

--_c0158b0f-e0e0-4d70-9531-283d71becc77_
Content-Type: text/html; charset="iso-8859-1"
Content-Transfer-Encoding: quoted-printable

<html>
<head>
<style><!--
.hmmessage P
{
margin:0px=3B
padding:0px
}
body.hmmessage
{
font-size: 10pt=3B
font-family:Tahoma
}
--></style></head>
<body class=3D'hmmessage'><div dir=3D'ltr'>
<div>Hi=2C In our project to use the HIVE on CDH3U4 release (Hive 0.7.1)=2C=
 I have a hive table like the following:</div><div><br></div><div>Table foo=
 (</div><div>&nbsp=3B &nbsp=3Bsearch_results array&lt=3Bstruct&lt=3Bid:bigi=
nt=2C..............&gt=3B&gt=3B</div><div>&nbsp=3B &nbsp=3Bsearch_clicks ar=
ray&lt=3Bstruct&lt=3Bindex_loc:int=2C ..............&gt=3B&gt=3B</div><div>=
)</div><div><br></div><div>As you can see=2C the 2nd column=2C which repres=
ents the list of search results clicked=2C contains the index location of w=
hich results was clicked (starting from 0). Now I need to find out the clic=
k count of some IDs=2C but this IDs only exists in the search_result struct=
.</div><div><br></div><div>When I tried to write a query for this=2C I will=
 try to do following first:</div><div><br></div><div><br></div><div>select =
search_results[c.index_loc] from foo lateral view explode(search_clicks) cl=
ickTable as c=3B</div><div><br></div><div>But it won't work in hive=2C as t=
he following error message returned:</div><div>AILED: Error in semantic ana=
lysis: Line 1:7 Non-constant expressions for array indexes not supported ra=
nk.</div><div><br></div><div>It looks like that currently Hive (at least 0.=
7) does NOT support non-constant expressions as the array index.</div><div>=
<br></div><div>I searched on google=2C and found out the following HIVE jir=
a ticket:</div><div><br></div><div><a href=3D"https://issues.apache.org/jir=
a/browse/HIVE-1955">https://issues.apache.org/jira/browse/HIVE-1955</a></di=
v><div><br></div><div>It looks like someone has the same request to support=
 it=2C but not yet.</div><div><br></div><div>But there is a comment in the =
above ticket that it can be done in an UDF. My question is=2C can anyone sh=
are some ideas about how to archive this in an UDF=2C as it maybe the only =
option for me right now?</div><div><br></div><div>Thanks</div> 		 	   		  <=
/div></body>
</html>=

--_c0158b0f-e0e0-4d70-9531-283d71becc77_--